Fix the deletion of SD3 text encoders for Dreambooth/LoRA training if the text encoders are not being trained #8536
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update. |
sayakpaul
added a commit
that referenced
this pull request
Dec 23, 2024
… the text encoders are not being trained (#8536) * Update train_dreambooth_sd3.py to fix TE garbage collection * Update train_dreambooth_lora_sd3.py to fix TE garbage collection --------- Co-authored-by: Kashif Rasul <[email protected]> Co-authored-by: Sayak Paul <[email protected]>
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
What does this PR do?
Fixes #8511 (issue)
If text encoder training is not enabled, and the text encoders are not needed during training, they are deleted, and garbage collection/cuda cache clear are run to free up the memory. Previously the deletion only included the lists referencing the text encoders and tokenizers, but because they are also referenced directly in their own variables, the referenced models weren't actually being deleted. This fix is a crude attempt to delete all references to the text encoders, so they can actually be removed from memory.
Before submitting
documentation guidelines, and
here are tips on formatting docstrings.
Who can review?
Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.
@sayakpaul